GNSS Spoofing Detection using Machine Learning

Abstract

Global navigation satellite systems (GNSS) and related electronic technologies are becoming increasingly paramount in environmental, engineering, and navigation contexts. Even so, radio frequency (RF) communication can interfere with civilian GNSS transmissions. The main aim is to make a GNSS receiver acquire and track false navigational information. Phase, energy, and fictitious signal components are used to analyze the differences between spoofing and real signal patterns. The correlation output of a tracking loop is utilized to derive three essential parameters, namely early-late phase, delta, and signal level, which are crucial in the signal extraction process. Machine learning techniques, such as K Nearest Neighbors, Neural Networks, and Naive Bayes classifiers are used for spoofing detection.GNSS spoofing detection uses a multilayer neural network with feature index inputs. Simulation results with a software GNSS receiver show that the ANN can achieve sufficient detection accuracy in a short juncture.